†Department of Informatics and Mathematical Modeling

Abstract-Microfluidic biochips offer a promising alternative to a conventional biochemical laboratory. There are two technologies for the microfluidic biochips: droplet-based and flow-based. In this paper we are interested in flow-based microfluidic biochips, where the liquid flows continuously through pre-defined micro-channels using valves and pumps. We present an approach to the system-level modeling and simulation of a cell culture microfluidic biochip called ProCell, Programmable Cell Culture Chip. ProCell contains a cell culture chamber, which is envisioned to run 256 simultaneous experiments (viewed as a 16 x 16 matrix). We use an inverted fluorescence microscope to observe the experiments in real-time, allowing kinetic data analysis. We are able to automati-cally adjust the current experimental setup thus allowing, for the first time, conditional experiments. We propose a biochip architecture model and a comprehensive fault model that captures permanent faults occurring during chip operation. Using the proposed modeling and simulation framework, we perform an architectural level evaluation of two cell culture chamber implementations. A qualitative success metric is also proposed to evaluate chip performance in the presence of partial failures. Our results show that significant improve-ments in efficiency can be obtained using redundancy, provid-ing improved chances to complete an experiment even in the presence of faults. This decreases the experiment repetition rate while increasing system productivity, saving time and reducing costs. I